**** BEGIN LOGGING AT Sat Aug 30 02:59:57 2008 Aug 30 06:21:47 03  07org.openembedded.dev * r17a4225dab 10OE.dev/conf/distro/include/preferred-om-2008-versions.inc: preferred-om-2008-versions.inc: Bumped bluez-hcidump to 1.42 Aug 30 06:21:54 03  07master * r17a4225dab 10OE.dev/conf/distro/include/preferred-om-2008-versions.inc: preferred-om-2008-versions.inc: Bumped bluez-hcidump to 1.42 Aug 30 06:29:14 03  07org.openembedded.dev * r82827c4f15 10OE.dev/packages/bluez/ (3 files): bluez: Updated rest of the packages to 3.33 (since 3.28 no longer exists as a recipe). Aug 30 06:29:17 03  07master * r82827c4f15 10OE.dev/packages/bluez/ (3 files): bluez: Updated rest of the packages to 3.33 (since 3.28 no longer exists as a recipe). Aug 30 07:20:57 morning all Aug 30 07:21:30 morning RP Aug 30 07:21:41 RP: a question for you Aug 30 07:22:06 when using virtual/kernel, and your PREFERRED kernel fails to build, why does bitbake then go and build every other kernel under the sun? Aug 30 07:22:20 (and is there a way to stop that happening) Aug 30 07:26:29 rwhitby: That is the traditional bitbake behaviour Aug 30 07:26:47 rwhitby: I don't actually agree with it but changing it would breakbacwards compatibilty Aug 30 07:27:04 I guess it was once a feature but its turned out no to be so useful Aug 30 07:27:34 We should make that behaviour configurable. It only happens with the -k option anyway Aug 30 07:48:03 RP: what was the use case in which it was useful, when PREFERRED_PROVIDER was defined? Aug 30 07:48:33 rwhitby: The idea was it would allow builds to complete rather than fail Aug 30 07:48:35 I can understand it if PREFERRED_PROVIDER is not defined, but cannot think of a good reason why bitbake should go against explicit directions for a PREFERRED_PROVIDER Aug 30 07:49:01 rwhitby: It comes does to the variable name - "PREFERRED" Aug 30 07:49:25 Would it be useful to change the design, I say yes Aug 30 07:49:53 but there is history there and as the maintainer I need to be careful with this Aug 30 07:49:58 understood Aug 30 07:50:15 can we make it configurable and configure to not do it by default ? Aug 30 07:51:23 We can make it configurable Aug 30 07:51:31 I would want to see a discussion about defaults Aug 30 07:51:48 Feel free to start that on the bitbake+OE list Aug 30 07:51:55 I'm actually in favour of changing it Aug 30 07:52:05 but it must be discussed Aug 30 08:07:14 nod Aug 30 08:12:35 RP: RFC sent to oe-devel Aug 30 08:13:44 rwhitby: and the bitbake list? Aug 30 08:24:16 03  07org.openembedded.dev * rd64b374096 10OE.dev/packages/tasks/task-openmoko-feed.bb: task-openmoko-feed: Added pidgin Aug 30 08:24:26 03  07master * rd64b374096 10OE.dev/packages/tasks/task-openmoko-feed.bb: task-openmoko-feed: Added pidgin Aug 30 09:06:25 hi all Aug 30 09:06:34 steliosk: ping Aug 30 09:06:48 likewise : hi ! Aug 30 09:06:58 steliosk: hi, long time no see here. Aug 30 09:07:15 steliosk: I was fixing up ffmpeg as well. What failed in install / strip? Aug 30 09:07:34 heh the joys of family and long work hours... Aug 30 09:08:25 likewise : it was trying to use the host not OE build one, result was it was failing for ppc Aug 30 09:08:58 steliosk: yes because it used "-s" on the install command. Aug 30 09:09:11 steliosk: I think the proper fix is to remove the "-s" Aug 30 09:09:34 steliosk: because we won't want the Makefile to strip the executables/libs for us. Aug 30 09:09:41 ...don't... Aug 30 09:10:09 steliosk: and you forgot to add "libavformat-makefile.patch" to the repo. Aug 30 09:10:58 hold on phone.... Aug 30 09:14:44 ok i am back... Aug 30 09:15:17 likewise : let me check about the patch... Aug 30 09:15:36 steliosk: please first read my fix proposal above Aug 30 09:17:01 likewise : i am not sure which way is the OE-way... i have no problem patching the Makefile and removing striping, bu on the other hand we will have to hunt/look every makefile for every ffmpeg version that comes out Aug 30 09:17:38 steliosk: the thing is, we want to package unstripped version of the library if someone is building a DEBUG build Aug 30 09:17:55 hmm correct Aug 30 09:18:01 steliosk: OE takes care of stripping itself, conditionally Aug 30 09:18:22 steliosk: I know the problem will be next releases. Aug 30 09:18:47 steliosk: alternatively, we could use "sed" to remove all "install -s" instances in the Makefile by "install" Aug 30 09:20:17 its oj by me Aug 30 09:20:24 s/oj/OK Aug 30 09:20:39 steliosk: I just wanted to know if you fixed anything else in the two missing patches? Aug 30 09:21:39 probably something that make ffmpeg not build on ppc :) but can't remember its like a week or so. I'll have a look at the patches Aug 30 09:22:28 likewise : BTW i am using our own distro which is not in sync with .dev and moving changed back and forth Aug 30 09:22:49 probably this is why i forgot the patches Aug 30 09:22:51 steliosk: ok, what are you targetting mostly? Aug 30 09:23:06 You're working with PowerPC's much? Aug 30 09:23:07 likewise : clients :) Aug 30 09:23:12 cool :-) Aug 30 09:23:20 going well? Aug 30 09:23:29 mostly ppc these days and some x86 Aug 30 09:23:45 Used AMCC PPC460EX? Aug 30 09:24:02 http://www.pikatechnologies.com/english/View.asp?x=652 Aug 30 09:25:04 steliosk: nice exposure Aug 30 09:28:38 steliosk: OK, I just pushed out the fix. You may want to test drive it. Aug 30 09:28:57 likewise : missing patch -> http://rafb.net/p/jRq6Pv90.html Aug 30 09:29:33 steliosk: OK, that one I missed. What happened without it? Aug 30 09:31:36 I'll add it. Looks sane Aug 30 09:31:42 likewise : does ffmpef build for you without this ? Aug 30 09:31:54 yes, it builds for avr32 without that patch. Aug 30 09:32:27 avr32 beeing uclibc.... hmmm Aug 30 09:32:38 steliosk: yes. what where you using when it failed? Aug 30 09:32:45 glibc Aug 30 09:32:49 gcc 4.1.1 Aug 30 09:33:14 for ppc? Aug 30 09:33:33 ^H^H^H powerpc :-) Aug 30 09:33:37 yes Aug 30 09:35:40 03  07org.openembedded.dev * re5a73ee64d 10OE.dev/packages/ (5 files in 4 dirs): ffmpeg: Revert DEPENDS on coreutils-native. Instead, remove -s (strip) argument to install. Aug 30 09:35:48 03  07master * re5a73ee64d 10OE.dev/packages/ (5 files in 4 dirs): ffmpeg: Revert DEPENDS on coreutils-native. Instead, remove -s (strip) argument to install. Aug 30 09:36:11 likewise : this is what happens without the patch -> http://rafb.net/p/u7VdhS24.html Aug 30 09:36:52 FWIW, removing the -s is the right thing to do Aug 30 09:37:16 install's strip functionality is rather broken anyway... Aug 30 09:40:01 RP: thanks, I figured removing "-s" is sane simply because we also want to build debug libs. Didn't know install's strip is rather broken. Does install respect env variable STRIP btw? Aug 30 09:42:50 likewise: No, it doesn't Aug 30 09:43:04 which is one of the ways its broken Aug 30 09:43:16 -s is wrong for the debug reasons you mention Aug 30 09:43:31 RP: that one is fixed now. Aug 30 09:43:42 steliosk: will commit your link fix now. Aug 30 09:43:50 thanks both Aug 30 09:48:23 03  07org.openembedded.dev * r8f731b50a4 10OE.dev/packages/ffmpeg/ (2 files in 2 dirs): ffmpeg: Brought in ifaistos' missing patch for fixed linker arguments for libavformat. Needed for ppc/glibc. Aug 30 09:48:31 03  07master * r8f731b50a4 10OE.dev/packages/ffmpeg/ (2 files in 2 dirs): ffmpeg: Brought in ifaistos' missing patch for fixed linker arguments for libavformat. Needed for ppc/glibc. Aug 30 09:49:00 likewise : still getting errors http://rafb.net/p/G8UHvo33.html Aug 30 09:49:58 steliosk: your link patch got in in r7, the strip patch only was r6 Aug 30 09:50:14 steliosk: see the CIA-3 commit above Aug 30 09:50:30 steliosk: please pull, clean and build again :-) Aug 30 09:50:32 hmm there is another patch missing Aug 30 09:51:13 steliosk: r7 is my final commit. If it still missing a fix, I'll leave it to you for now :-) Aug 30 09:57:05 likewise : one other thing i have changed is the SECTION from libs to console/multimedia. any objections ? Aug 30 09:57:34 steliosk: no sounds good to me Aug 30 09:58:19 FWIW, I'm trying to build the GNU Flash player for AVR32, along with ffmpeg for the media support now. Aug 30 10:00:03 RP: Can we expect better x86 support soonish? :-) Aug 30 10:36:30 03  07org.openembedded.dev * rbc622730c0 10OE.dev/packages/ffmpeg/ (2 files in 2 dirs): packages/ffmpeg/ffmpeg_0.4.9-pre1.bb : Add libmp3lame if mp3 support is configured in Aug 30 10:36:36 03  07master * rbc622730c0 10OE.dev/packages/ffmpeg/ (2 files in 2 dirs): packages/ffmpeg/ffmpeg_0.4.9-pre1.bb : Add libmp3lame if mp3 support is configured in Aug 30 10:42:05 * * OE Bug 4200 has been UNCONFIRMED by vtzan(AT)teilam.gr Aug 30 10:42:07 * * test bug after moving from Greg's server to amethyst Aug 30 10:42:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4200 Aug 30 10:55:03 RP: I have a package foo that was build for arch AAA, so packages are foo.AAA.ipk .Then i added some machine specific MMM files anf the package name changed to foo.MMM.ipk. Can i force an existing image to drop the arch packet and use the machine specific one ? Aug 30 11:21:34 steliosk: what do you mean by an "existing image"? Are you asking whether you can do an online upgrade with ipkg to switch the packages from one architecture to another? Aug 30 11:21:36 hi lrg Aug 30 11:36:35 pb__ yes Aug 30 11:37:02 steliosk: I think that will work. You just need to make sure that the new architecture is a higher priority. Aug 30 11:42:40 pb__ That seems to do it Aug 30 11:43:09 ok, very good Aug 30 11:44:09 pb__ For some reason i am getting the machine with 2 priorities.... Aug 30 11:44:51 03  07org.openembedded.dev * r600128e78b 10OE.dev/packages/tasks/task-openmoko-feed.bb: task-openmoko-feed: added epdfview Aug 30 11:44:59 03  07master * r600128e78b 10OE.dev/packages/tasks/task-openmoko-feed.bb: task-openmoko-feed: added epdfview Aug 30 11:45:02 pb__ http://rafb.net/p/k77pA147.html machine is warp Aug 30 11:45:26 that is a bit odd. what is ${PACKAGE_ARCHS}? Aug 30 11:46:32 oh, it has ${TARGET_ARCH} in there by default, that's probably your problem Aug 30 11:46:43 PACKAGE_ARCHS = "all any noarch ${TARGET_ARCH} ${PACKAGE_EXTRA_ARCHS} ${MACHINE}" Aug 30 11:46:44 yea itsdefined there also, that's why Aug 30 11:47:27 yeah. I guess you could make package_generate_ipkg_conf() a bit smarter so that it filters out those duplicate entries Aug 30 11:47:52 or just redefine PACKAGE_ARCHS to not mention ${TARGET_ARCH} Aug 30 11:48:20 * pb__ bbiab Aug 30 11:48:22 pb__ for the moment i will follow the path of least...work ;) Aug 30 11:50:20 re Aug 30 11:50:34 hey folks Aug 30 11:58:42 fascinating... when I google "openembedded pocket loox 720" it is clear that google understands that openembedded = OE Aug 30 13:10:00 how bad is using definition like char var[strlen(other_var)]; ? Aug 30 13:10:04 in C Aug 30 13:30:12 not bad at all, why should it be? Aug 30 13:30:38 of course, if you plan to copy other_var into this one, you'd better add an extra byte for the terminator, but the definition itself is fine Aug 30 13:35:10 pb__: is such behavior standartiesed? Aug 30 13:35:22 I think it only C++.. Aug 30 13:36:05 yes, it's been standard C for nearly a decade Aug 30 13:36:21 hm.. Aug 30 13:36:42 I've rewrite it to malloc() already :) Aug 30 13:36:59 but now I know :) Aug 30 13:37:28 heh Aug 30 13:46:49 anyone else having trouble updating from monotone.openembedded.net atm, or is it just me? Aug 30 13:49:03 ah, nm, it's just being slow Aug 30 13:50:34 bluelightning: yes I have trouble it being slow :-) Aug 30 13:51:53 How do I specify the version correctly? using "bitbake ffmpeg-0.4.9+svnr12186" gives me NoProvider: ffmpeg-0.4.9+svnr12186 Aug 30 13:54:02 likewise: maybe just set PREFERRED_VERSION_ffmpeg Aug 30 13:54:07 in your local.conf Aug 30 13:55:14 bluelightning: tnx, I will fall back to that approach now. Aug 30 14:48:06 * Jay7 is away up to tomorrows evening.. Aug 30 15:40:47 | + log_check rootfs Aug 30 15:40:47 | + set +x Aug 30 15:40:47 | log_check: Using /diskb/openembedded/build/tmp/work/ks8695-angstrom-linux-gnueabi/minimal-image-1.0-r0/temp/log.do_rootfs.19854 as logfile Aug 30 15:40:47 | log_check: There were error messages in the logfile Aug 30 15:40:47 | log_check: Matched keyword: [ERR] Aug 30 15:40:49 | (offline root mode: not running util-linux-umount.postinst) Aug 30 15:40:51 | Collected errors: Aug 30 15:40:53 | * ERROR: Cannot satisfy the following dependencies for task-boot: Aug 30 15:40:55 | * glibc (>= 2.6.1) * Aug 30 15:40:57 | Aug 30 15:40:59 NOTE: Task failed: /diskb/openembedded/build/tmp/work/ks8695-angstrom-linux-gnueabi/minimal-image-1.0-r0/temp/log.do_rootfs.19854 Aug 30 15:41:02 NOTE: package minimal-image-1.0-r0: task do_rootfs: failed Aug 30 15:41:12 anybody can help me? Aug 30 16:09:53 03  07org.openembedded.dev * rb91aaa1b71 10OE.dev/packages/libqpe/ (libqpe-opie/citytime-path-2.patch libqpe-opie_cvs.bb): Aug 30 16:09:53 libqpe-opie: fix error in do_patch compilation for cvs version Aug 30 16:09:53 * Add citytime-path-2.patch which applies cleanly against current CVS Aug 30 16:09:53 03  07org.openembedded.dev * r86b42ef16c 10OE.dev/packages/opie-security/opie-security_cvs.bb: opie-security: install ownerinfo.png in cvs version Aug 30 16:09:55 03  07org.openembedded.dev * r43d29e419d 10OE.dev/packages/opie-cardapplet/ (opie-cardapplet.inc opie-cardapplet_cvs.bb): Aug 30 16:09:58 opie-cardapplet: add detection of card insert/remove and fix postinst Aug 30 16:10:00 * Add udev scripts/rules to notify applet of card insertion/removal Aug 30 16:10:02 * Fix path to qcop used in postinst Aug 30 16:10:04 03  07master * r43d29e419d 10OE.dev/packages/opie-cardapplet/ (opie-cardapplet.inc opie-cardapplet_cvs.bb): Aug 30 16:10:09 opie-cardapplet: add detection of card insert/remove and fix postinst Aug 30 16:10:11 * Add udev scripts/rules to notify applet of card insertion/removal Aug 30 16:10:13 * Fix path to qcop used in postinst Aug 30 16:10:15 03  07master * rb91aaa1b71 10OE.dev/packages/libqpe/ (libqpe-opie/citytime-path-2.patch libqpe-opie_cvs.bb): Aug 30 16:10:18 libqpe-opie: fix error in do_patch compilation for cvs version Aug 30 16:10:20 * Add citytime-path-2.patch which applies cleanly against current CVS Aug 30 16:10:24 03  07master * r86b42ef16c 10OE.dev/packages/opie-security/opie-security_cvs.bb: opie-security: install ownerinfo.png in cvs version Aug 30 17:11:20 03  07org.openembedded.dev * r3fec0c7c50 10OE.dev/packages/opie-clockapplet/ (3 files): opie-clockapplet: install alarm icon image Aug 30 17:11:26 03  07org.openembedded.dev * r4d0cb3adf0 10OE.dev/packages/opie-pics/ (opie-pics.inc opie-pics_1.2.3.bb opie-pics_cvs.bb): opie-pics: install symlink icon for file selector Aug 30 17:11:27 03  07master * r3fec0c7c50 10OE.dev/packages/opie-clockapplet/ (3 files): opie-clockapplet: install alarm icon image Aug 30 17:11:28 03  07master * r4d0cb3adf0 10OE.dev/packages/opie-pics/ (opie-pics.inc opie-pics_1.2.3.bb opie-pics_cvs.bb): opie-pics: install symlink icon for file selector Aug 30 21:41:30 hi all Aug 30 21:41:34 hi flo_lap Aug 30 21:45:10 03  07org.openembedded.dev * r85b1b183c8 10OE.dev/packages/ (12 files in 7 dirs): Fix paths to qcop Aug 30 21:45:16 03  07org.openembedded.dev * re61d1f0a25 10OE.dev/packages/ (9 files in 5 dirs): opie-mediaplayer1: Add 1.2.3 versions of mediaplayer1 and plugins Aug 30 21:45:16 03  07org.openembedded.dev * rb238cc4d45 10OE.dev/packages/ (4 files in 2 dirs): Fix path to qcop Aug 30 21:45:17 03  07master * r85b1b183c8 10OE.dev/packages/ (12 files in 7 dirs): Fix paths to qcop Aug 30 21:45:18 03  07master * rb238cc4d45 10OE.dev/packages/ (4 files in 2 dirs): Fix path to qcop Aug 30 21:45:21 03  07master * re61d1f0a25 10OE.dev/packages/ (9 files in 5 dirs): opie-mediaplayer1: Add 1.2.3 versions of mediaplayer1 and plugins Aug 30 22:52:12 03  07org.openembedded.dev * ra1196fe844 10OE.dev/packages/portabase/metakit_2.4.9.3.bb: metakit: fix source URL Aug 30 22:52:17 03  07master * r71f9cf988d 10OE.dev/packages/qpealarmclock/qpealarmclockapplet_1.0.9.bb: Fix path to qcop Aug 30 22:52:18 03  07master * ra1196fe844 10OE.dev/packages/portabase/metakit_2.4.9.3.bb: metakit: fix source URL Aug 30 22:52:18 03  07master * r2504f1532b 10OE.dev/packages/zshopi/zshopi_0.2.bb: zshopi: fix source URL Aug 30 22:52:19 03  07org.openembedded.dev * r71f9cf988d 10OE.dev/packages/qpealarmclock/qpealarmclockapplet_1.0.9.bb: Fix path to qcop Aug 30 22:52:22 03  07org.openembedded.dev * r2504f1532b 10OE.dev/packages/zshopi/zshopi_0.2.bb: zshopi: fix source URL **** ENDING LOGGING AT Sun Aug 31 02:59:56 2008